Heart valve disease, left ventricular hypertrophy, and heart failure: a lifelong relationship and continuing clinical responsibility

This article refers to ‘Heart failure with preserved ejection fraction after left-sided valve surgery: prevalent and relevant’ by A.A. Kammerlander et al., published in this issue on pages 2008–2016. It is abundantly clear that clinical outcomes after interventions for left-sided heart valve disease are related to preoperative left ventricular hypertrophy (LVH) and myocardial fibrosis. The timing of treatment is influenced by the detection of changes in left ventricular (LV) geometry and long-axis function and pulmonary vascular resistance, with the goal of avoiding persistent symptoms and irreversible progression to heart failure. It is important to diagnose adverse remodelling preoperatively and to assess its consequences for LV function and reserve, and also important to monitor responses after valve surgery to confirm that hypertrophy is regressing and that function is maintained or improved. Kammerlander et al.1 report that two thirds of their patients attending for follow-up on average 50 months after surgery for aortic or mitral valve disease, fulfilled diagnostic criteria for heart failure with preserved ejection fraction (HFpEF). More than half of the patients were in atrial fibrillation and their natriuretic peptide levels were very high, exceeding levels usually observed in patients with HFpEF alone. More importantly, only 12.5% of them had been diagnosed to have heart failure, so their symptoms and need for increased treatment were unrecognized. And most importantly, patients with this syndrome had an increased mortality that was similar to patients with heart failure with reduced ejection fraction (HFrEF), who had been recognized. Over a follow-up of 7 years their risk of dying was 1.8 times higher than patients who had also had valve interventions but who did not have heart failure. Their study serves as a timely reminder that full recovery after valve surgery cannot be assumed, so careful assessment and continuing close supervision are indicated. Aortic stenosis (AS) is mostly a disease of the elderly. Hypertension is common in older people. Hypertension causes LVH. Many patients with AS develop concentric LVH as a compensatory response to their outflow obstruction, that tends to normalize wall stress, but they may also be hypertensive. In the SEAS study of 1616 patients, hypertension was associated with a 56% higher rate of ischaemic cardiovascular events and a two-fold increased mortality, independently of the severity of AS or of abnormal LV geometry.2 Some patients who have a normal LV ejection fraction and are diagnosed to have normal-flow low-gradient AS may have predominant hypertensive heart disease.3 Even some patients diagnosed to have low-flow low-gradient AS may instead have a combination of HFpEF with mild or moderate AS. It can be difficult to discriminate between these since both may impair stroke volume reserve, but hypertension with a wide pulse pressure indicative of stiff central conduit arteries and impaired ventricular–arterial coupling is at least a contributory cause of LVH in many patients. Volume and pressure overload from aortic regurgitation lead to diastolic dysfunction with increased myocardial stiffness that persists late after valve surgery.4 Elderly patients with severe mitral regurgitation may have coexisting hypertension that contributes to the progression of disease and development of eccentric LVH. Chronic LV dilatation may exacerbate secondary mitral regurgitation. Thus there are many overlaps between the patterns of LV and left atrial (LA) remodelling caused by hypertensive heart disease and by heart valve disease, with particular patterns for each pathology. The most recent European Society of Cardiology (ESC) consensus recommendations stated that “HFpEF can be suspected if there is … no significant heart valve disease” and that “HFpEF ‘masqueraders’ such as heart valve disease … need to be excluded.”5 Heart valve disease affects not only LV mass but also most of the other echocardiographic measurements that contribute to the HFA-PEFF score. For example, the septal e′ may be reduced if there is interstitial fibrosis in AS, or increased if there is volume overload caused by aortic or mitral regurgitation. The E/e′ ratio may be useful in patients with AS but unreliable in those with mitral valve disease.6 Long-axis function (measurable as LV global longitudinal strain, another component of the score) can be reduced after cardiac surgery because of pericardial adhesions. Atrial fibrillation secondary to LA enlargement, irrespective of whether or not there is coincidental heart failure, contributes to high levels of natriuretic peptides. Natriuretic peptide levels are also increased after cardiac surgery. In mitral stenosis the left atrium is enlarged and there may be post-capillary pulmonary hypertension, both changes that can also occur in HFpEF. Thus, it is inappropriate either to describe patients as having ‘HFpEF’ due to heart valve disease, or to diagnose that a patient with valve disease has coexisting HFpEF without first undertaking a careful search for specific haemodynamic sequelae of surgery. That does not invalidate the observation that patients with valve disease may have an equivalent phenotype — with LVH, diastolic dysfunction, longitudinal systolic dysfunction, impaired LV functional reserve, LA dilatation, and secondary pulmonary hypertension — that is equally deserving of effective treatment. Unfortunately, Kammerlander et al. did not have data to report preoperative findings in their patients either about the severity of AS or other valve diseases or about the degree of LVH, but it is more likely that the phenotype at follow-up was persisting rather than new. The investigators applied diagnostic criteria from the ESC guidelines of 2017, which for HFpEF were rather general. The precise diagnostic label, however, is less important than recognition of its components and their consequences — especially hypertrophy. After aortic valve replacement (AVR) there is less regression of LVH in patients who have poorly controlled hypertension, and persisting LVH predicts poor outcomes including heart failure. Earlier observations from the surgical literature have been confirmed after transcatheter aortic valve implantation, for example in 1434 patients from the PARTNER trials in whom less regression of LVH at 1 year was associated with increased mortality.7 The risk ratio for those with severe LVH at 1.77 was almost identical to that reported by Kammerlander et al. for their ‘HFpEF’ phenotype,1 so in essence their study reinforces the known prognostic significance of hypertension, LVH, atrial fibrillation, LA volume, and natriuretic peptide levels after valve surgery. Risk factors for the persistence of LVH include a body mass index ≥30 kg/m2 before AVR8 and either metabolic syndrome or diabetes.9 There is less regression if the ascending aorta is heavily calcified10 or if postoperative renal function is moderately impaired.11 Jang et al.12 reported that LVH correlated most closely with global LV afterload and valvuloarterial impedance. It is notable that all these factors are also risk factors for HFpEF5; a common mechanism would be that they all increase arterial stiffness and augment end-systolic loading. The use of warfarin after mechanical valve replacement is also relevant: since it inhibits the matrix Gla protein, which is itself an inhibitor of vascular calcification, it can increase arterial stiffness and pulse pressure13 which would promote LVH. A secondary benefit from optimal control of hypertension after valve interventions could be a reduction in systemic embolism. After AVR, postoperative hypertension had an odds ratio of 8 for ischaemic cerebrovascular events, while other risk factors included postoperative smoking and diabetes.14 In the Cardiff Embolic Risk Factor Study, which enrolled patients after all heart valve operations, hypertension and diabetes contributed to a score that strongly predicted systemic embolism.15 Replacement fibrosis within the LV myocardium is probably irreversible but interstitial fibrosis can regress after AVR.16 Diffuse fibrosis as a consequence of chronic maladaptive LVH correlates with diastolic dysfunction. Relative expansion of the extracellular volume fraction, as a marker of diffuse fibrosis, appears more pronounced in women than in men at comparable degrees of pressure overload.17 Patients who undergo valve interventions are typically elderly with multiple comorbidities. They may have hypertension, atherosclerotic diseases and arrhythmias as well as heart valve disease, so clinical responsibilities do not end with the intervention (Figure 1). In the study by Kammerlander et al., overall mortality after 83 ± 39 months was 34%.1 As well as individual treatment directed to the primary target of hypertension with LVH, there is a need for more evidence to guide therapeutic decisions. Since most patients are now enrolled in registries, there is scope for large registry-based randomized trials. With a factorial or adaptive design they could evaluate both established and new drugs that are considered for patients with HFpEF. Well-characterized age-matched control groups would be essential. These comments have mostly concerned AS, which was the dominant lesion in 62% of the patients in the study of Kammerlander et al. In a randomized study of 114 patients after AVR, candesartan reduced LV mass and LA volume and increased LV long-axis systolic function; the changes were unrelated to systolic blood pressure or pulse pressure which were similar to controls.18 In an observational study of 200 patients with LVH who received a bioprosthetic AVR, the probability of complete LV mass regression by 1 year postoperatively was increased by an odds ratio of 3.4 in patients who had been prescribed a beta-blocker or calcium-channel blocker.19 In the OCEAN-TAVI registry of 1215 subjects, blockade of the renin–angiotensin system was associated with greater reduction of LV mass and a lower mortality at 2 years (by >50%).20 Until more evidence is available from prospective trials about the efficacy of specific pharmacological interventions, the main conclusion that may be derived from the study by Kammerlander et al. is that we should continue to search for more precise diagnostic criteria that would qualify patients for surgery early enough to prevent irreversible adverse sequelae of preoperative LVH and myocardial fibrosis, and which would justify the risks of an earlier surgical procedure. Better diagnostic criteria will be specific for each valve lesion, and all patients may have classical HFpEF without it being caused by their valve disease. Conflict of interest: none declared.